echodot

joined 2 years ago
[–] echodot@feddit.uk 1 points 11 hours ago (1 children)

If we have colonies on Mars I can't see why we also wouldn't have colonies on the moon if only to support ships going further out into the solar system. Refuelling on the moon makes much more sense than refuelling on earth.

The only way that isn't the case is if we build a space elevator to bring resources up to spacecraft orbit, but if we're at that point I still think we would build colonies on the moon just because, even if it's just at the level of a Saudi Arabian vanity project.

[–] echodot@feddit.uk 1 points 12 hours ago

Would we be able to see it from earth?

I've seen pictures of the earth from the moon and you can barely identify the continents. You definitely can't see individual open cast mines.

[–] echodot@feddit.uk 1 points 12 hours ago

Ok but walk it back a bit, why did they become homeless?

If somebody is completely 100% mentally healthy I can't see how an AI can convince them to kill themselves any more than another person could convince them to kill themselves. Only vulnerable people join cults, because it's difficult to pray on people who have proper defences.

I'm still not convinced that the AI isn't just triggering some underlying mental condition that other people in their lives are just not aware of or not willing to accept.

[–] echodot@feddit.uk 1 points 12 hours ago

Some people think that LLMs are true AGI or at least they have thoughts that run along those lines even if they can't articulate it like that.

They tend to be people who aren't particularly tech savvy and so they see this thing that seems to be pretty much a miracle of technology and believe that it truly is a super intelligence.

I've seen evolution simulators come up with some truly interesting behaviour, like finding shortcut glitches in Mario that no human has ever found, if I didn't know how the program worked I suppose I might believe that there was some intelligence there.

[–] echodot@feddit.uk 1 points 12 hours ago

I'd had a negative opinion of Asimov's laws of robotics being used to control AI for most of my life, and LaMDA successfully persuaded me to change my opinion.

Then he's an idiot.

Asimov's laws of robotics aren't some kind of model by which to control AI, there are plot device. They're literally not supposed to work, if they did work it would be a very short book, so obviously we shouldn't use them for controlling AI.

I don't know any serious IT professional that has ever, at any point, ever forwarded the opinion that an AI (should we ever a create one, because there is an arguement that LLMs aren't AI) should be ruled by a plot device from a book. Equally if we ever invent warp drive and find aliens I'm assuming we're not going to be restricted to the prime directive.

[–] echodot@feddit.uk 1 points 12 hours ago

I think the important point here is that just because the father is doing Google doesn't necessarily mean that Google are at fault. People tend to feel that if an individual is suing a corporation for malfeasance the corporation is necessarily guilty. But reality doesn't always run like that.

I can't see any reason that Google would want to encourage more suicide so I have to assume that it's just an unfortunate interaction of a mentally unsound mind and a product that frankly even its own creators don't understand. This is highly unfortunate but I'm not certain where the crime was.

[–] echodot@feddit.uk 1 points 12 hours ago

Yes people can have mental delusions and psychotic episodes; I'm not necessarily convinced that they are a separate unique condition simply because they were triggered by an AI versus anything else.

For one thing I've yet to hear a decent (or indeed any) explanation as to the mechanism by which AI triggers psychosis that is materially different from any other trigger. Most people who suffer from this condition can be triggered by literally anything, including mundane things such as seeing a red cars slightly more often than they believe they should, then they concoct this conspiracy about an evil cabal of red car owners.

[–] echodot@feddit.uk 1 points 13 hours ago* (last edited 13 hours ago)

A little bit alarmist I feel, after all if it was this easy to be affected by AI about half the population would be dead by now, so clearly it's not that simple.

[–] echodot@feddit.uk 1 points 14 hours ago

The collective of all human comments on the topic distilled together are greater than that of Andrew Tate. Simply because Andrew Tate goes out of his way to be an awful person. Clueless is better than evil.

[–] echodot@feddit.uk 22 points 2 days ago

My bank once sent me a letter to my address, to tell me that they did not know what my address was. So I'm not completely sure they are exactly on the ball.

[–] echodot@feddit.uk 2 points 2 days ago (1 children)

Binary for three. Two would be 1 0 which would then get confused with 10 which would actually take us full circle.

[–] echodot@feddit.uk 37 points 2 days ago

I think it's incel for having a shower.

view more: next ›